At present, services integrated into a multiprocessor core system are increasing gradually, and network I/O (input/output, input/output) requests are also increasing. The data processing capability of the multiprocessor core system to network I/O requests plays a key role for the system to exert its full capability. In the prior art, one method for processing network I/O requests is: based on a static-configuration processing mode, binding a task to a designated processor core, where the task is performed on the designated processor core, or assigning the tasks on idle processor cores in a balanced manner, where the tasks are performed on the idle processor cores; another method for processing network I/O requests is: based on a static-configuration switchover rule, switching over the processor core between an interruption mode and a polling mode to improve the processing efficiency of the system to the network I/O requests.
During implementation of the foregoing technical solution, the inventor finds that the prior art faces at least the following challenges: static configuration refers to manual configuration prior to the running process of the system, and during the actual running process of the system, since a processor core is scheduled based on the static configuration solution, the demand for real-time network I/O processing cannot be met and a reasonable solution of scheduling a processor core cannot be provided, thereby failing to improve the efficiency of the overall multiprocessor core system.